PROJECT SUMMARY Brain activity never ceases. When we are asleep, inattentive, or even under general anesthesia, networks of interconnected neurons in the human brain continue to spontaneously generate complex activity patterns. Sensory stimuli perturb this ongoing spontaneous neuronal activity. In order to be consciously detected, the effect of this perturbation needs to be large enough so as to engage thousands of neurons and persist for at least several hundred milliseconds. When we are awake and attentive, the smallest stimuli are sufficient to elicit a large perturbation. Under general anesthesia, however, even the most noxious stimuli do not reach the threshold for conscious perception. Here we address a fundamental question: why are sensory stimuli able to perturb neuronal activity in some states but not in others? We hypothesize that the ability of the sensory stimuli to perturb neuronal activity is related to the property of dynamical systems termed stability. If neuronal dynamics were unstable, the effect of any perturbation would grow over time without bounds and engage ever increasing number of neurons. Conversely, if the dynamics were too stable, then all perturbations will quickly dampen down and fail to reach threshold of perception. Thus, we hypothesize that conscious perception is most likely to occur when the neuronal dynamics are poised precisely between the stable and unstable regimes. We refer to this point as critical. To test the criticality hypothesis, we developed novel mathematical techniques and applied them to neurophysiological recordings in humans and in nonhuman primates. These preliminary findings strongly support the hypothesis. In the proposed project, we will rigorously test the criticality hypothesis using electrocorticography (ECoG) in human subjects implanted with electrodes for epilepsy localization. We will determine how the stability of spontaneous activity varies as a function of sleep and wake, attentiveness and drowsiness, as well as sedation and general anesthesia. We will validate the criticality hypothesis and our ability to estimate stability of neuronal activity by predicting responses to electrical brain stimulation. Using an auditory masked speech detection task, we will also determine whether stability of neuronal dynamics can be used to predict whether a natural stimulus presented at perceptual threshold will be consciously detected. While many other measures of neuronal activity have been previously associated with changes in arousal and perception, at present, it is not possible to apply the existing measures to unequivocally distinguish between activity in the conscious and unconscious brain. Hence, validating this criticality hypothesis would be a major advance. In addition to addressing a fundamental issue in neuroscience, finding an objective and quantifiable measure of sensory responsiveness has profound clinical significance in neurology and in anesthesiology where diagnoses of covert awareness under anesthesia or after brain injury cannot be made reliably with existing technology.